Systems and methods for improving bit rate control of video encoding by correcting for the effects of scene changes and bit stuffing in a video buffer verifier (VBV) buffer model

1. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison; said scale factors for the set summing to 1;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor, different from the first factor and normalized to sum, in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped.

2. The method as defined in claim 1 , further comprising:

detecting that a scene change has occurred; and

invoking the sum computing step, the comparing step, and one of the reassigning steps only upon the detection of the scene change.

3. The method as defined in claim 1 , further comprising:

detecting that the video encoding process is in a critical state; and

invoking the sum computing step, the comparing step, and one of the reassigning steps only upon the detection of the scene change.

4. The method as defined in claim 1 , wherein the first state of the comparison comprises a determination that sum is less than the predetermined threshold T norm .

5. The method as defined in claim 1 , further including the step of computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer.

6. The method as defined in claim 1 , wherein a value for the predetermined threshold T norm is zero.

7. The method as defined in claim 1 , wherein the video encoding process corresponds to MPEG-2.

8. The method as defined in claim 1 , wherein the video encoding process is performed in real time.

9. The method as defined in claim 1 , wherein the video encoding process is performed in real time and with a constant bit rate.

10. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder at an effectively constant average bit rate;

detecting that the video encoding process is in a critical state; and

invoking the sum computing step, the comparing step, and one of the reassigning steps only upon the detection of the scene change;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped; wherein the critical state is identified when bit stuffing is used, but values for quantization parameter mquant j are above a predetermined threshold.

11. The method as defined in claim 10 , wherein the predetermined threshold is a lowest value for the quantization parameter permitted by a corresponding encoding standard.

12. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder at an effectively constant average bit rate;

computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped; wherein delta comprises:

delta= vbv _buffer_size− V ststus ,

where vbv_buffer_size is a capacity of a model of the buffer at the decoder and V status is a current buffer occupancy of the model of the buffer at the decoder.

13. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder at an effectively constant average bit rate;

computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped; wherein delta comprises:

delta= V initial −V status ,

where V initial is an initial value of a capacity of a model of the buffer at the decoder and V status is a current buffer occupancy of the model of the buffer at the decoder.

14. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder at an effectively constant average bit rate;

computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped; wherein assigning virtual buffer levels according to the first criterion further comprises:

d 0 i =delta· fr i ;

d 0 p =delta· fr p ; and

d 0 b =delta·fr b ,

where d 0 i , d 0 p and d 0 b are the previously assigned virtual buffer occupancy initial values for the I-pictures, the P-pictures and the B-pictures, respectively, fr i , fr p and fr b are the first scaling factors for the I-pictures, the P-pictures and the B-pictures, respectively.

15. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for I-pictures, for P-pictures and for B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning each of the virtual buffer occupancy initial values to a value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture;

computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer;

wherein assigning virtual buffer levels according to the first criterion further comprises:

d 0 i =delta· fr i ;

d 0 p =delta· fr p ; and

d 0 b =delta· fr b ,

where d 0 i , d 0 p and d 0 b are the previously assigned virtual buffer occupancy initial values for the I-pictures, the P-pictures and the B-pictures, respectively, fr i , fr p and fr b are the first scaling factors for the I-pictures, the P-pictures and the B-pictures, respectively;

wherein values for fr i , fr p , and fr b correspond to 5/17, 5/17, and 7/17, respectively.

16. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

comparing the value for sum to a predetermined threshold T norm ;

reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison;

reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor in response to a second state of the comparison; and

providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder at an effectively constant average bit rate;

computing a value for delta, where delta is the value corresponding to an unoccupied amount of the decoder buffer;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped; wherein assigning virtual buffer levels according to the second criterion further comprises:

d

0

i

=

d

0

i

·

delta

sum

;

d

0

p

=

d

0

p

⁣

·

delta

sum

;

⁢

and

d

0

b

=

d

0

b

⁣

·

delta

sum

,

where d 0 i , d 0 p and d 0 b are the previously assigned virtual buffer occupancy initial values for the I-pictures, the P-pictures and the B-pictures, respectively.

17. A computer readable medium encoded with computer executable instructions for reassigning virtual buffer levels to stabilize a video encoding process, comprising:

instructions for assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

instructions for computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

instructions for comparing the value for sum to a predetermined threshold T norm ;

instructions for reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled respectively by a corresponding first factor in response to a first state of the comparison; said scale factors for the set summing to 1;

instructions for reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor, different from the first factor and normalized to sum, in response to a second state of the comparison; and

instructions for providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped.

18. A circuit for reassigning virtual buffer levels to stabilize a video encoding process in a real-time video encoder comprising:

means for assigning at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

means for computing a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

means for comparing the value for sum to a predetermined threshold T norm ;

means for reassigning to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled by a first factor in response to a first state of the comparison; said scale factors for the set summing to 1;

means for reassigning each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor, different from the first factor and normalized to sum, in response to a second state of the comparison; and

means for providing the reassigned virtual buffer occupancy initial values to the video encoding process prior to encoding a next picture for encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate;

wherein said set of I-pictures, P-pictures and B-pictures are not skipped.

19. A circuit adapted to reassign virtual buffer levels to stabilize a video encoding process comprising:

a module configured to assign at an encoder initial values for virtual buffer occupancies for a set of I-pictures, P-pictures and B-pictures, respectively, each of the virtual buffer occupancies indicative of an occupancy level of a buffer at a decoder for each of a corresponding one of the I-pictures, the P-pictures and the B-pictures, respectively, prior to encoding a first block of a picture;

a module configured to compute a value for sum, where sum corresponds to a summation of the virtual buffer occupancy initial values for the I-pictures, for the P-pictures, and for the B-pictures;

a module configured to compare the value for sum to a predetermined threshold T norm ;

a module configured to reassign each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to an adaptively modified value corresponding to an unoccupied amount of the decoder buffer scaled by a first factor in response to a first state of the comparison; said scale factors for the set summing to 1; and

a module configured to reassign to alter each of the virtual buffer occupancy initial values for said set of I-pictures, P-pictures and B-pictures to respective previously assigned values thereof scaled by a second factor, different from the first factor and normalized to sum, in response to a second state of the comparison;

wherein encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer occurs at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate; and said set of I-pictures, P-pictures and B-pictures are not skipped.

20. A video encoder that embodies the circuit of claim 19 .

21. A method of reassigning virtual buffer levels to stabilize a video encoding process, the method comprising:

monitoring a final occupancy level d of a virtual buffer for a set of I-pictures, P-pictures and B-pictures, the virtual buffer representing occupancy of a physical buffer at a decoder;

determining whether the final occupancy level d corresponds to a value impossible to realize in the physical buffer; and

resetting to alter the final occupancy level d for said set of I-pictures, P-pictures and B-pictures upon a positive determination that the value is impossible to realize in the physical buffer;

wherein encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer occurs at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate, and said set of I-pictures, P-pictures and B-pictures are optionally not skipped.

22. The method as defined in claim 21 , wherein the final occupancy level d of the virtual buffer corresponds to the type of picture that had just been encoded by the video encoding process.

23. The method as defined in claim 21 , wherein determining comprises comparing the final occupancy level to zero such that a value less than zero indicates a final occupancy level that is impossible to realize in the physical circuit.

24. The method as defined in claim 21 , wherein resetting comprises resetting to zero.

25. The method as defined in claim 21 , further comprising normalizing values for the final occupancy levels of other picture types when the final occupancy level d is reset.

26. A computer readable medium encoded with computer executable instructions for reassigning virtual buffer levels to stabilize a video encoding process, comprising:

instructions for monitoring a final occupancy level d of a virtual buffer for a set of I-pictures, P-pictures and B-pictures, the virtual buffer representing occupancy of a physical buffer at a decoder;

instructions for determining whether the final occupancy level d corresponds to a value impossible to realize in the physical buffer; and

instructions for resetting to alter the final occupancy level d for said set of I-pictures, P-pictures and B-pictures upon a positive determination that the value is impossible to realize in the physical buffer;

wherein encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer occurs at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate; and said set of I-pictures, P-pictures and B-pictures are optionally not skipped.

27. A circuit for reassigning virtual buffer levels to stabilize a video encoding process in a real-time video encoder comprising:

means for monitoring a final occupancy level d of a virtual buffer for a set of I-pictures, P-pictures and B-pictures, the virtual buffer representing occupancy of a physical buffer at a decoder;

means for determining whether the final occupancy level d corresponds to a value impossible to realize in the physical buffer; and

means for resetting to alter the final occupancy level d for said set of I-pictures, P-pictures and B-pictures upon a positive determination that the value is impossible to realize in the physical buffer;

wherein encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer occurs at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate; and said set of I-pictures, P-pictures and B-pictures are optionally not skipped.

28. A circuit adapted to reassign virtual buffer levels to stabilize a video encoding process comprising:

a module configured to monitor a final occupancy level d of a virtual buffer for a set of I-pictures, P-pictures and B-pictures, the virtual buffer representing occupancy of a physical buffer at a decoder;

a module configured to determine whether the final occupancy level d corresponds to a value impossible to realize in the physical buffer; and

a module configured to reset to alter the final occupancy level d for said set of I-pictures, P-pictures and B-pictures upon a positive determination indicates that the value is impossible to realize in the physical buffer;

wherein encoded transmission of said set of I-pictures, P-pictures, and B-pictures to the decoder buffer occurs at an effectively constant average bit rate; said set exchanged between the encoder and encoder buffer, and decoder buffer and decoder respectively, at an optionally variable bit rate; and said set of I-pictures, P-pictures and B-pictures are optionally not skipped.

29. A video encoder that embodies the circuit of claim 28 .